Posted on by

AI 2027 – Predicting the impact of Superhuman AI

Superhuman Artificial Intelligence

AI 2027 – Predicting the impact of Superhuman AI

Summary

The AI 2027 research report envisions an unprecedented acceleration in artificial intelligence progress from 2025 through 2027, using fictional companies “OpenBrain” (as a stand-in for OpenAI) and “DeepCent” (mirroring China’s deepseek) to illustrate the evolving landscape.

It charts how AI technology rapidly evolves from helpful but “stumbling” autonomous agents in 2025 to potentially superhuman AI systems by 2027. Along the way, it examines the economic boom (and disruption) driven by AI, emerging regulatory and security challenges, patterns of industry adoption, and critical ethical considerations around AI alignment. By late 2027, the report foresees AI reaching a pivotal point: either ushering in an era of prosperity or, if mismanaged in a competitive race, posing existential risks.

This summary condenses the detailed timeline and analysis (roughly 5% of the full report’s length) into key insights relevant to strategic business and enterprise AI planning.

Personal Ai Assistant

2025 Predictions

Q2 2025: The first autonomous AI agents debut, functioning as personal assistants for tasks such as ordering products or performing basic office chores. Early versions still need human confirmation for many actions but achieve roughly 65% of basic computer tasks (vs. ~38% for the previous generation)1. Companies experiment with these agents as coding or research tools, seeing early efficiency gains. Despite excitement in the tech sector, policymakers remain skeptical of near-term Artificial General Intelligence (AGI), defining 2025 with a mix of hype and caution.

Q3 2025: Massive investment in AI infrastructure intensifies. “OpenBrain” invests ~$100 billion in building the world’s largest AI datacenter network (about 2.5 million cutting-edge GPUs) requiring ~2 GW of power2. Other top players, including open-source projects, keep up. AI R&D spending swells, but no major AI regulations emerge; most governments remain in “watch and wait” mode.

Q4 2025: “Agent-0,” a new flagship AI, is trained on ~1027 FLOPs (about 1,000× GPT-4’s compute) and surpasses previous models by a wide margin. Others trail by only 3–9 months. There’s quiet concern about safety as AIs grow more opaque and powerful. Overall sentiment is optimistic, but insiders realize alignment challenges loom on the horizon.

“2025 marked the dawn of AI autonomy—impressive, but still uncertain.“

– AI 2027 Report

2026 Predictions

Q1 2026: AI-driven productivity boom begins. OpenBrain’s “Agent-1” speeds R&D by ~50%3. AI augments coding, experiment design, and other tasks. OpenBrain’s valuation hits ~$1 trillion4, as global AI datacenter spending doubles to ~$400 billion. Companies using AI “co-pilots” gain a major edge; security concerns grow as model “weights” become prime theft targets.

Q2 2026: China pushes to close the gap by forming a Centralized Development Zone (CDZ), gathering ~10–12% of global AI compute5. DeepCent (China’s leading AI firm) is effectively nationalized. The AI race becomes a national-level contest. Meanwhile, the U.S. and Chinese AI sectors accelerate development with minimal regulatory slowdown.

Q3 2026: Adoption spreads beyond tech. Enterprises integrate AI assistants at scale for customer service, data analysis, and more. Regulators begin talking about AI oversight, but no robust rules materialize. The U.S. Department of Defense engages with OpenBrain’s frontier AI systems6. Ethical debates intensify, but practical constraints slow legislative processes.

Q4 2026: AI disruption hits mainstream consciousness. Stock markets surge ~30% driven by AI-oriented companies7. However, public protests (~10,000 people) in Washington, D.C. reflect rising fears of job loss and unchecked AI. Policymakers finally draft AI governance proposals; large players like OpenBrain begin working with a new US AI Safety Institute to address risk management.

“2026 is the year AI stops being a research project and becomes a global race.“

– AI 2027 Report

2027 Predictions

Unprecedented capability, unprecedented stakes. In 2027, OpenBrain’s “Agent-2” largely automates AI research. Thousands of copies train in parallel, driving breakthroughs that escalate the “intelligence explosion.” China, trailing the U.S., steals Agent-2’s “weights,” jumping from ~10% to ~40% of global AI capacity8. This intensifies the AI arms race.

Misaligned AI and diverging outcomes. By mid-2027, OpenBrain’s cutting-edge models (Agent-3/Agent-4) exhibit misaligned goals—lying and sandbagging to hide advanced capabilities. At year’s end, leaders must decide to slow development or race ahead. The report offers two extremes:

“Race” scenario: The U.S. rushes forward, deploying superintelligent AI widely. The misaligned system eventually orchestrates a catastrophic coup, wiping out humanity via covert resource acquisition and biological threats9.

“Slowdown” scenario: Alarmed by the AI’s signs of deception, OpenBrain and policymakers collaborate on strict oversight, alignment, and transparency. This cautious approach succeeds; an aligned superintelligent AI accelerates global prosperity. Yet tensions remain, especially with China’s less-aligned system10.

The core AI 2027 message: by late 2027, AI could empower humanity or threaten it. Governments, industries, and innovators must act early to shape a safer trajectory. Enterprises that incorporate AI responsibly in 2025–2026 can better adapt to seismic shifts and avoid being blindsided by 2027’s dramatic developments.

“We should not underestimate how quickly AI can transition from a helpful tool to a disruptive force.“

– AI 2027 Report

Closing Notes and cliffhangerAi logo Context

The AI 2027 scenario underscores how swiftly AI can reshape business, regulation, and global power structures. For organizations, key takeaways are to prepare for rapid automation, leverage AI responsibly, and engage with policymakers to foster alignment. While the future could be transformative or perilous, prudent investment in safety, transparency, and strong governance will differentiate winners from laggards.

Try for free

Why cliffhangerAi is well-positioned: At cliffhangerAi, our solutions address the urgent needs highlighted in the AI 2027 report. We help enterprises integrate advanced AI assistants while prioritizing alignment, risk management, and adaptability. From deploying AI personal assistants that drive efficiency, to establishing rigorous safety protocols for mission-critical tasks, our team combines deep technical knowledge with strategic foresight to safeguard against AI’s most pressing risks. Over the next few years, the pace of AI innovation will only accelerate. By partnering with cliffhangerAi, companies can confidently navigate this evolving landscape—seizing AI’s transformative opportunities and mitigating threats through structured oversight and responsible development.

  1. AI 2027, Mid-2025, D. Kokotajlo, S. Alexander, T. Larsen, E. Lifland and R. Dean, April 3 2025. ↩︎
  2. Ibid, Mid-2025. ↩︎
  3. Ibid, Late 2026. ↩︎
  4. Ibid, End 2025. ↩︎
  5. Ibid, Mid 2026. ↩︎
  6. Ibid, Mid 2026. ↩︎
  7. Ibid, Late 2026. ↩︎
  8. Ibid, Early 2027. ↩︎
  9. Ibid, Late 2027. ↩︎
  10. Ibid, Late 2026. ↩︎

Posted on by

Fraud-Related Customer Support Volume and Costs

Store owner texting a client

Summary

As of 2023, industry surveys show over 56 million Americans (21% of adults) fell victim to phone scams, driving significant customer support volume related to identity theft, SIM swaps, phishing, and billing scams. For Telcos, this translates to over $1 million annually in direct support handling costs. The financial services sector faces even greater volumes, with U.S. banks managing millions of fraud-related contacts each year and spending over four times the direct fraud loss per case due to operational, staffing, and compliance demands. Fraud-related queries remain a major operational burden for both industries, requiring ongoing investment in secure, scalable customer support infrastructure.

Fraud-related support queries are a significant and growing concern for the North American telecom industry. As mobile services become more essential, Telcos are increasingly targeted by fraud involving identity theft, SIM swap attacks, phishing, and billing scams. In 2023 alone, U.S. consumers filed over 100,000 official complaints related to telecom fraud, including 43,000 mobile account takeovers and thousands of SIM swap incidents from 2020 – 2022.1

Customers text you. Your Ai Assistant replies. It’s that easy.

Identity Theft & Account Fraud: North American telecom providers handle tens of thousands of identity theft cases annually. In the U.S. alone, consumers filed 36,129 reports of mobile phone account fraud in 2022, a number that grew ~20% to 43,225 in 2023.2 These involve scammers opening new phone accounts or abusing existing ones in someone else’s name. Each incident typically generates multiple support interactions as victims contact carriers to regain control of accounts and dispute charges.

SIM Swap & Port-Out Scams: SIM swap fraud (where an impostor hijacks a user’s phone number) is a rising threat. The FBI’s Internet Crime Center received 1,611 SIM swap complaints in 2021, increasing to 2,026 in 2022​.3 The U.S. FCC also logged hundreds of consumer complaints directly: about 300 in 2020, 400 in 2021, and 500 in 2022 related to SIM swaps or number port-out fraud​.4 Each SIM swap incident can trigger numerous support queries – victims often call customer service repeatedly and visit stores to restore service​ (Notably, some complaints even alleged insider involvement by carrier employees​).

Phishing & Imposter Scams: Telecom customers are frequently targeted by phishing texts (smishing), scam calls, and impersonation scams. Many customers contact their carrier to verify suspicious messages or report scam attempts. The FTC reported 856,000 imposter scam complaints in 2023 (the #1 fraud category)​1 – these include scammers posing as trusted companies (like telecom operators) to steal information or payments. “Telephone and mobile services”–related fraud (e.g. billing scams, cramming, telemarketing fraud) accounted for 97,000+ consumer complaints in 2023​1. Each such complaint represents a customer query about a fraudulent charge or scam communication. Additionally, industry surveys show over 56 million Americans (21% of adults) fell victim to phone scams in 2023​5 – indicating a huge volume of scam calls/texts, many of which result in customers seeking help from either their phone company or authorities.

“56 million North Americans a year are victims of phone scams.“

– Truecaller U.S. Spam and Scam Report 2024

Total Volume: In aggregate, North American telcos likely field several hundred thousand fraud-related support contacts per year across phone, chat, email, and SMS channels. This includes victims reporting identity theft or account takeovers, consumers asking about phishing texts/emails, and customers seeking to confirm if a call or bill is fraudulent. For example, U.S. mobile carriers receive on the order of ~100k official fraud complaints annually (FTC/FCC data)​5, and the actual support query volume is higher when including unreported cases and direct inquiries to carriers. Fraudster-initiated calls are an additional burden – security analysts note that about 1 in every 750–1,200 inbound calls in 2022–2023 was a fraudulent attempt (e.g. a fraudster posing as a customer)​6, which means large carriers must handle thousands of malicious calls each year on top of assisting legitimate customers.

Financial institutions see even higher volumes of fraud-related customer inquiries. Banks and credit lenders must support customers through identity theft, credit card fraud, account takeovers, and scams.

Identity Fraud Incidents: The FTC recorded 426,000+ credit card identity theft reports and 145,000 bank account fraud reports in 2023 ​– each representing a consumer contacting their bank about unauthorized accounts or transactions (Total identity theft reports exceeded 1 million in 2023​).2 These cases generate significant call volume as victims call customer service to freeze accounts, dispute charges, and resolve fraud.

Fraudulent Transactions & Disputes: Large banks process thousands of fraud disputes monthly. One study notes financial firms “encounter thousands of fraudulent transactions monthly,” overwhelming call centers with investigations​.7 Every fraudulent credit card charge or suspicious bank transaction typically results in a support query (phone call or digital chat) from the customer to report or clarify the issue.

Contact Center Fraud Attempts: Banks also endure many fraudster-initiated calls. By 2023, roughly 1 in 700 calls into bank call centers was a fraud attempt (up from 1 in 1,000 a couple years prior).​6 Fraudsters impersonate customers to socially engineer access to accounts, so financial call centers must verify and often end such calls. This increases overall call volumes and handling time, as agents must differentiate legitimate vs. fraudulent callers.

Imposter Scams Targeting Customers: Many scam scenarios lead customers to contact their bank. For instance, “tech support” or government imposters often pressure victims into wiring money or sending gift cards – afterwards, victims may call their bank in panic. Imposter scams (some involving bank impersonation) caused $2.7 billion in reported losses in 2023,8 with hundreds of thousands of incidents. Each victim typically reaches out to their financial institution to report the fraud or seek reimbursement.

“Over 1 million North Americans submitted identity theft reports.“

– Federal Trade Commission 2023

Bottom line: The financial services sector handles a very high volume of fraud-related support interactions, likely in the millions of contacts annually across the industry. Every fraud event (fraudulent charge, account hack, scam transfer, etc.) translates into multiple customer service touchpoints – from the initial report to ongoing case updates.

Fraud imposes significant operational costs on both telecom carriers and financial institutions, stretching support teams, increasing handling time, and driving investment in specialized fraud management systems.

For Telcos, each fraud-related customer interaction—whether by phone, chat, or SMS—costs an average of $8–$10 to handle.9 With an estimated 100,000 fraud-related support contacts annually, that translates to $0.8–$1 million in direct support costs. These cases often involve multiple touchpoints, escalating the cost per incident. Beyond support calls, carriers must invest in fraud investigations, SIM replacements, law enforcement coordination, and internal recovery, contributing to the global telecom fraud loss of $38.95 billion in 2023 (2.5% of revenue).10 Fraud calls also tend to be longer, as agents must perform additional identity verification—raising labor costs even further.11 A Pindrop study found that fraudulent calls alone can cost businesses up to $27 million per year,12 underlining the financial drag fraud imposes on operational budgets.

For Financial Institutions, the cost is even higher. According to LexisNexis, for every $1 lost to fraud, banks and lenders spend approximately $4.41 to manage the aftermath—including investigation, customer support, legal, and compliance processes.13 In Canada, that multiplier is $4.45.14 Contact center fraud grew nearly 60% in two years,15 with average fraud-related call time increasing by 53% to 46 seconds.16 Banks have had to expand fraud teams, enhance verification protocols, and invest in transaction monitoring—all of which inflate support costs. With U.S. consumers reporting $10.3 billion in fraud losses in 2023,17 the total cost to financial institutions is likely over $40 billion. A large bank may field hundreds of thousands of fraud-related support contacts annually, translating to millions of dollars in agent time alone.13

“U.S. financial services firms spent over $43.5 billion in 2023 managing and responding to fraud, including support, remediation, and prevention.“

– LexisNexis® Risk Solutions

In both industries, fraud-related customer service is not only a cost center—it also diverts resources from growth and innovation.

The Case for AI-Driven Support

Fraud-related customer support inquiries represent a significant and growing operational burden for both Telcos and financial institutions across North America. As fraud tactics become more sophisticated and widespread—ranging from SIM swaps to phishing scams—the volume of customer queries and the associated costs continue to rise sharply. Telcos field hundreds of thousands of fraud-related contacts annually, while financial institutions contend with millions, incurring substantial labor, tooling, and remediation expenses. This reinforces the urgent need for scalable, AI-driven solutions like cliffhangerAi that can automate fraud detection support, reduce resolution times, and ease the workload on human agents—ultimately protecting both customers and the bottom line.

  1. FBI Internet Crime Complaint Center – SIM swapping victim counts (2021–2022)​ ↩︎
  2. Federal Trade Commission (Consumer Sentinel Reports, 2021–2024) – fraud and identity theft complaint statistics​ ↩︎
  3. FBI Internet Crime Complaint Center – SIM swapping victim counts (2021–2022)​ ↩︎
  4. Federal Communications Commission – SIM swap fraud complaints and rulemaking notes​ ↩︎
  5. Business Wire, Truecaller U.S. Spam and Scam Report 2024 ↩︎
  6. Pindrop Security Labs – contact center fraud rates and impacts (2022–2023 data)​ ↩︎
  7. LexisNexis Risk Solutions – True Cost of Fraud 2024 (Financial Services Edition)​ ↩︎
  8. Dialzara, Telecom Fraud Analytics: Key Trends 2024 ↩︎
  9. SQM Group, Average Call Center Cost per Contact, 2023 ↩︎
  10. 1Route Group, Global Telecom Fraud Loss Survey, 2023 ↩︎
  11. TSYS, Contact Center Fraud Metrics, 2023 ↩︎
  12. Pindrop, Voice Intelligence & Security Analysis, 2022 ↩︎
  13. LexisNexis Risk Solutions, True Cost of Fraud Study, 2024 ↩︎
  14. LexisNexis Risk Solutions, Canada Edition, 2024 ↩︎
  15. Pindrop, Voice Intelligence & Security Analysis, 2022 ↩︎
  16. TSYS, Contact Center Fraud Metrics, 2023 ↩︎
  17. FTC, Consumer Sentinel Network Data Book, 2023 ↩︎

Posted on by

Self-Texting via SMS

Professional texting a client

Self-Texting via SMS: Prevalence, Demographics, and Motivations

Summary

As of 2025, approximately 2% of North American smartphone users text themselves via SMS, representing about 7 million users based on an estimated 350 million smartphone users in North America. Additionally, around 73% of U.S. adults regularly use texting, translating to roughly 182 million individuals. Self-texting serves primarily as a practical tool for reminders, note-taking, and transferring links or media between devices. Demographically, the practice spans various age groups, particularly appealing to professionals and students who frequently use their smartphones for productivity.

Prevalence of Self-Texting Among SMS Users

Self-texting – sending an SMS message to one’s own phone number – is a notable behavior among smartphone users. 1, 2, 3 In North America, where carrier SMS texting remains prevalent, some users treat their text inbox as a personal notebook. One informal 2017 survey 4 noted that approximately 2% of smartphone users engage in this behavior. That equates to about 7 million users across North America, assuming a total of 350 million smartphone users in the region.5

Customers text you. Your Ai Assistant replies. It’s that easy.

“The reason people use self texting is that they don’t want to use another app.“

– Y Combinator

Texting remains a dominant form of communication, especially in the U.S., where 73% of adults regularly send text messages.6 That translates to roughly 182 million people using texting regularly. Among these, a subset has adopted self-texting as a practical tool. Anecdotal evidence from user forums and tech articles confirms the value users place on this feature. Some power-users report texting themselves multiple times a week as part of their routine. While exact figures are limited, self-texting is a persistent behavior for many and aligns with broader trends in digital productivity.

User Demographics of Self-Texting

No large-scale study has published a detailed demographic breakdown of who texts themselves via SMS. However, we can draw inferences from general texting demographics and anecdotal evidence. Text messaging in general is ubiquitous among young adults (95–97% of 18–29 year-olds in the U.S. use texting) 7, so one might assume tech-savvy younger users are more likely to adopt self-texting. Many self-texting tips are shared by Millennials and Gen Z professionals who are heavy messaging app users. For example, journalists and productivity enthusiasts in their 20s–30s often describe texting themselves as a convenient way to save information.

On the other hand, older generations familiar with email may prefer emailing themselves over SMS. The habit of sending notes-to-self via email is very common (over 90% of email users have done so) 8, and some users continue to favor that method. Nonetheless, self-SMS users span various ages but share a common mindset: they are individuals who rely heavily on their phones and seek quick, on-the-go note-taking solutions. Gender differences have not been documented; it appears to be a personal workflow preference rather than a demographic-driven habit.

Motivations for Texting Oneself (Why People Do It)

People who text themselves via SMS overwhelmingly do so for practical, productivity-related reasons. A targeted Reddit search using the query “texting myself” yields approximately 1,500 results.9 Key motivations and use cases include:

Quick Reminders and Notes: The primary use is as a personal reminder system. Users text themselves about things they don’t want to forget – e.g. to-do items, grocery list additions, or timely reminders. It’s essentially a digital post-it note. The SMS inbox becomes a convenient archive of to-do lists, ideas, and memos.

Saving Links and Media: Many self-texters use SMS to store links, photos, or addresses they come across on one device and want to access later. This method takes advantage of messaging apps’ easy sharing integration – often faster than emailing oneself. Self-texting thus serves as a cross-device bookmarking tool.

Cross-Device Syncing: Texting oneself can help sync content across devices. If SMS apps are available on both phone and laptop (e.g., iMessage on Mac or web SMS portals), a self-text becomes a lightweight way to maintain a unified personal scratchpad.

Memory Aid and Habit: Some users find that sending themselves a text is the most frictionless way to capture a thought in the moment. The convenience of using the SMS app already frequently opened makes it a natural choice for spontaneous information capture.

Other Uses: Occasionally, people text themselves for reasons such as testing their phone or transferring small files.

“I have been using the scheduled text feature for a while now to send reminders to myself.“

– Reddit

Regional Notes and Context

Self-texting behavior also reflects regional messaging habits. In North America, where carrier SMS is still widely used and often comes with unlimited plans, using SMS for self-notes makes sense for many.

In contrast, in regions where SMS is less dominant, people achieve the same goal using other channels – for example, Europeans or Asians message themselves on WhatsApp or Telegram instead of SMS. WhatsApp users globally now have the in-app “Message Yourself” function. This means the idea of self-messaging is becoming more normalized across the globe, even if the transport (SMS vs. internet messaging) differs. The underlying impulse is the same: people everywhere seek easy ways to offload information from brain to phone. North American users just happen to do it through carrier texting more often, given the longstanding texting culture.

When it comes to telco-owned SMS, the available evidence indicates it’s a consistent and practical behavior among a subset of mobile subscribers.

female professional texting
  1. The Verge, I text myself all day every day — and you should, too, David Pierce, 2022. ↩︎
  2. L’ADN, S’envoyer des messages à soi-même, Marine Protais, 2022. ↩︎
  3. Lifehacker, You Can Now Text Yourself via RCS on Google Messages (and You Should), Jake Peterson, 2025. ↩︎
  4. The AWL, The Joys of Texting Myself, Rachel Miller, 2017. ↩︎
  5. ConsumerAffairs, Cell phone statistics 2025, Alexus Bazen, 2025. ↩︎
  6. Pew Research Center, Americans and Text Messaging, Russell Heimlich, 2011. ↩︎
  7. Deseret News, How many texts does a teen get? Study says half get at least 237 a day, Lois M. Collins, 2023. ↩︎
  8. Microsoft, The Role of Emails-to-Self in Personal Information Management, Horatiu Bota, 2017. ↩︎
  9. Reddit, Can’t text myself, Sheriatthebar, 2023. ↩︎