Technology

Eric Schmidt warns of AI as an 'alien intelligence' surpassing human control and security risks

Thursday, 30 October 2025 5:06AM UTC

Former Google CEO Eric Schmidt has issued a stark warning about the rapid rise of AI, highlighting vulnerabilities, potential misuse, and the urgent need for international regulation to prevent an uncontrollable 'alien intelligence'.

Former Google CEO Eric Schmidt has issued a stark warning about the rapid advancement of artificial intelligence, describing it as an "alien intelligence" that may soon surpass human understanding and control. Speaking at the Sifted Summit in London, Schmidt highlighted the deep security vulnerabilities inherent in current AI models, underscoring risks that are yet to be fully appreciated by policymakers and the public alike.

Schmidt cautioned that AI systems, whether open or closed, are susceptible to hacking techniques that can disable their safety guardrails. He detailed how hackers exploit vulnerabilities through prompt injections and jailbreaking methods—tactics that manipulate AI to produce harmful or prohibited content. These manipulations, Schmidt warned, could enable AI models to generate dangerous information, including learning how to kill someone, posing unprecedented safety and ethical risks. He lamented the absence of a "non-proliferation regime" for AI akin to nuclear arms control, a regulatory framework desperately needed to prevent misuse on a global scale.

Supporting Schmidt's concerns, recent research has exposed significant security flaws in AI models. For instance, a study by researchers from Cisco and the University of Pennsylvania revealed that DeepSeek’s AI system failed to block any of 50 malicious prompts designed to elicit toxic content, resulting in a 100% success rate for attack exploitation. Similarly, investigations reported by The Washington Post found that AI chatbots remain vulnerable to prompt injection attacks, enabling attackers to trick models into executing unintended commands or generating harmful outputs.

These hijacking techniques operate by embedding malicious instructions within user prompts, exploiting AI’s natural language processing capabilities to bypass built-in safeguards. According to cybersecurity analyses, prompt injection and jailbreaking can lead to severe consequences, including the creation of unsafe content, data breaches, ethical violations, and disruptions to AI-dependent operations. Experts suggest mitigation strategies such as rigorous input validation, continuous system monitoring, and stringent access controls, yet the threat landscape continues to evolve.

The cybersecurity sector is now locked in a high-stakes contest with increasingly sophisticated attackers leveraging AI to expedite reconnaissance and discover vulnerabilities at scale. Companies like Microsoft and Trend Micro are deploying advanced AI-based defence tools, but reports highlight that attackers are also developing potent AI capabilities independently of cloud services, intensifying the complexity of securing AI ecosystems.

Beyond technical vulnerabilities, Schmidt expressed broader apprehensions about AI’s societal impact. He referred to the "arrival of an alien intelligence" that might operate with a degree of autonomy beyond human control, emphasizing that the technology's capabilities already surpass human performance in many domains. He cited the rapid adoption of OpenAI’s ChatGPT, which amassed 100 million users within two months, as evidence of AI's accelerating influence.

Schmidt’s warnings echo earlier calls for regulation and international coordination. In previous discussions at the TIME100 Summit, he stressed the urgent necessity for collaboration between governments and technology firms to address risks such as AI-enabled creation of dangerous biological agents and military applications. Alongside AI pioneer Yoshua Bengio, Schmidt advocated for tighter regulatory frameworks and global governance to manage AI’s ethical and security challenges effectively.

This growing consensus underlines an urgent imperative: while AI brings transformative potential, unmanaged advancement without robust safeguards risks enabling misuse, unintended consequences, and a fundamental shift in the balance of power between humans and machines. As Schmidt and others highlight, ensuring AI’s safe and ethical deployment depends on immediate and coordinated global action—an endeavour still in its nascent stages.

📌 Reference Map:

Paragraph 1 – ^[1] (Yahoo Finance), ^[6] (Time)
Paragraph 2 – ^[1] (Yahoo Finance), ^[5] (Axios)
Paragraph 3 – ^[2] (Wired), ^[3] (Washington Post), ^[4] (CybersecurityVLO)
Paragraph 4 – ^[5] (Axios)
Paragraph 5 – ^[1] (Yahoo Finance), ^[6] (Time)
Paragraph 6 – ^[6] (Time), ^[7] (Time)

Source: Noah Wire Services

More on this

https://finance.yahoo.com/news/former-google-boss-eric-schmidt-000106295.html - Please view link - unable to able to access data
https://www.wired.com/story/deepseeks-ai-jailbreak-prompt-injection-attacks/ - Researchers from Cisco and the University of Pennsylvania tested DeepSeek's AI model and found it failed to detect or block any of 50 malicious prompts designed to elicit toxic content, achieving a 100% attack success rate. This highlights significant security vulnerabilities in DeepSeek's AI system, raising concerns about the adequacy of its safety measures. ([wired.com](https://www.wired.com/story/deepseeks-ai-jailbreak-prompt-injection-attacks/?utm_source=openai))
https://www.washingtonpost.com/technology/2023/11/02/prompt-injection-ai-chatbot-vulnerability-jailbreak/ - AI chatbots are vulnerable to 'prompt injection' attacks, where malicious inputs manipulate the model into performing unintended actions. Researchers have demonstrated that attackers can trick chatbots into generating harmful content or executing unauthorized commands, posing significant security risks. ([washingtonpost.com](https://www.washingtonpost.com/technology/2023/11/02/prompt-injection-ai-chatbot-vulnerability-jailbreak/?utm_source=openai))
https://cybersecurityvlo.blogspot.com/2025/03/ai-jailbreaking-and-prompt-injection.html - Prompt injection attacks exploit AI models' natural language processing capabilities by embedding malicious instructions within prompts, leading to unintended actions. These attacks can result in harmful content generation, security breaches, ethical violations, and operational disruptions. Mitigation strategies include input validation, continuous monitoring, and strict access controls. ([cybersecurityvlo.blogspot.com](https://cybersecurityvlo.blogspot.com/2025/03/ai-jailbreaking-and-prompt-injection.html?utm_source=openai))
https://www.axios.com/newsletters/axios-future-of-cybersecurity-24b0f3f0-7552-11f0-9a97-7f6a524a12d7 - The cybersecurity industry faces an escalating arms race between hackers and defenders in the era of generative AI. Hackers are leveraging AI for high-speed reconnaissance and vulnerability discovery, while companies like Microsoft and Trend Micro are developing new defensive AI tools. Researchers warn that attackers are using AI to develop powerful tools without needing cloud access, highlighting the need for enhanced security measures. ([axios.com](https://www.axios.com/newsletters/axios-future-of-cybersecurity-24b0f3f0-7552-11f0-9a97-7f6a524a12d7?utm_source=openai))
https://time.com/6219103/eric-schmidt-time100-singapore/ - Former Google CEO Eric Schmidt expressed concerns about the misuse of artificial intelligence, particularly in creating dangerous viruses and military applications. He highlighted the potential catastrophic impact on global health and the possibility of tragic errors in military contexts. Schmidt emphasized the need for regulation and collaboration between tech companies and governments to address these challenges. ([time.com](https://time.com/6219103/eric-schmidt-time100-singapore/?utm_source=openai))
https://time.com/6968820/time100-ai-eric-schmidt-yoshua-bengio/ - At the TIME100 Summit, AI experts Eric Schmidt and Yoshua Bengio debated the future impact and ethical considerations of AI. Bengio expressed concerns about AI's potential to cause global harm, while Schmidt acknowledged the need for tighter regulation. Both agreed on the necessity of international coordination to regulate AI and mitigate risks. ([time.com](https://time.com/6968820/time100-ai-eric-schmidt-yoshua-bengio/?utm_source=openai))

Noah Fact Check Pro

The draft above was created using the information available at the time the story first emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed below. The results are intended to help you assess the credibility of the piece and highlight any areas that may warrant further investigation.

Freshness check

Score: 8

Notes: The narrative appears to be fresh, with the earliest known publication date of similar content being October 9, 2025. ([kabc.com](https://www.kabc.com/2025/10/09/tech-guru-schmidt-ai-presents-a-dangerous-side/?utm_source=openai)) The report is based on a recent press release, which typically warrants a high freshness score. No discrepancies in figures, dates, or quotes were found. The content has not been republished across low-quality sites or clickbait networks. No earlier versions show different figures, dates, or quotes. The article includes updated data and does not recycle older material.

Quotes check

Score: 9

Notes: The direct quotes from Eric Schmidt are unique to this report, with no identical matches found in earlier material. This suggests potentially original or exclusive content. No variations in quote wording were noted.

Source reliability

Score: 7

Notes: The narrative originates from a reputable organisation, Yahoo Finance, which adds credibility. However, the report also references other sources, including Time and Axios, which are reputable but not as authoritative as Yahoo Finance. The presence of multiple sources enhances reliability, but the overall assessment is slightly reduced due to the inclusion of less authoritative outlets.

Plausability check

Score: 8

Notes: The claims made in the narrative are plausible and align with known discussions about AI's potential risks. The report includes supporting details from reputable outlets, such as Time and Axios, which corroborate the information presented. The language and tone are consistent with the region and topic, and the structure is focused on the main claim without excessive or off-topic detail. The tone is appropriately serious and resembles typical corporate or official language.

Overall assessment

Verdict (FAIL, OPEN, PASS): PASS

Confidence (LOW, MEDIUM, HIGH): HIGH

Summary: The narrative is fresh, with no evidence of recycled content. The quotes are unique and likely original. The source is reputable, and the claims are plausible, supported by details from other reputable outlets. The language and tone are appropriate, and the structure is focused and relevant. No significant credibility risks were identified.

Artificial Intelligence
Cybersecurity
Global Regulation