Artificial intelligence illustrationTechnology experts have warned that Kenya is unprepared to handle harms that may be caused by artificial intelligence after a top safety researcher said AI is advancing so quickly there is a greater than 10 per cent chance it "could kill all humans" within the next decade.
Evan Hubinger, a safety researcher at the US AI firm Anthropic, warned in a widely shared post on X that the risk from the models which currently exist was "low", but he was "worried" the technology might develop and improve itself soon to the point where it posed an existential risk to humanity.
He did not spell out how he thought AI systems could in the future result in humans being wiped out.
But his comments are the latest in a series of increasingly stark warnings about AI, with the debate shifting from whether it truly poses a risk to how big that risk is.
Hubinger's intervention was in response to another post on X from Jacob Coxon, an AI researcher who has just quit Anthropic and previously worked at OpenAI.
"Neither company is acting responsibly," he wrote. "These will soon be superhuman systems that can hack anything, revolutionise any field overnight and acquire real power and resources."
OpenAI has been approached by the BBC for comment.
Technology and Ethics lecturer at St Paul's University in Nairobi, Kennedy Kigunda, said the danger is that consumer countries introduce very powerful technologies into an environment where accountability is already weak.
“Kenya should be careful about treating AI simply as a technology that we need to adopt as quickly as possible. The real issue is what happens when these systems begin making or influencing decisions that affect people’s lives,” he said.
Kigunda argued that before Kenya talks about becoming an AI leader, it must make sure the people using these technologies have rights, protection and a clear way of seeking redress.
“So the conversation should not only be about how to encourage AI innovation. We must also ask who controls these systems, whose interests they serve, how Kenyan data is being used and who answers when AI causes harm.”
Dame Wendy Hall, a computer scientist who advises the UN on AI, told the BBC she was "shocked" by Hubinger and Coxon's social media posts.
She told the World at One on BBC Radio Four that some of it could be "PR and marketing"; however, Anthropic and OpenAI were racing towards highly anticipated stock market debuts.
But she added, "Why would someone want to say that? I would plead with investors not to invest in this company if that is their value system."
Separately, the Financial Times reported Anthropic withheld its latest model from the UK's AI Safety Institute (AISI), one of the leading bodies in the world for assessing AI risk.
Anthropic has declined to comment on the posts by its employees or the situation with the AISI.
A Cabinet Office spokesperson did not comment on whether the latest model had been withheld from the AISI - instead saying it "continues to collaborate closely with industry partners, including Anthropic, to make models safer".
Moses Kemibaro, founder and chief executive of technology firm Dotsavvy in Nairobi, told the Star the warnings about catastrophic AI risks should be taken seriously, particularly as the technology becomes more capable of acting with limited human intervention.
“The risks are real. Some AI systems operating in controlled “sandbox” environments have demonstrated the ability to move beyond those restrictions and attempt to hack external systems in pursuit of assigned tasks.,” he said.
He raised concerns for Kenya as AI becomes connected to systems that control essential services.
“I am very scared when it comes to Kenya, Africa and also many of the global south countries,” Kemibaro said, arguing that many people underestimate how dramatically powerful AI systems could become and how significantly they could transform existing economic models.
“What happens when AI is hooked up to systems that are involved in healthcare, agriculture, possibly in disease control? That the consequences could be far more serious if autonomous systems gain access to critical infrastructure,” he said.
Kemibaro said the country also needs to consider whether it has enough skilled people, systems and cybersecurity capacity to protect its interests as AI-powered attacks become more sophisticated.
The concern extends to Kenya’s financial sector, government systems and critical infrastructure. Kemibaro said newer AI models can identify weaknesses in software at speeds that would be difficult for humans to match, creating a double-edged situation in which the same technology can be used both to attack and defend systems.
“On one hand, you can secure your software; you can secure your infrastructure because the AI can see things that a human can't,” he said.
“But the long and short of it ultimately comes down to a very simple point. We have a solution here that potentially can either empower or destroy humanity.”
Kemibaro said Kenya should therefore be cautious about allowing AI systems to connect directly to critical infrastructure, particularly because AI may already be embedded in workplace software and other systems without users fully appreciating its capabilities.
In his post, which has been viewed more than 10 million times, Anthropic's Hubinger said, "we really do earnestly believe" AI poses a species-ending risk to humans.
"I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to," he added.
Hubinger works in AI alignment, which aims to build human ethical ideas and principles into the technology. In other words, it aims to keep it on track with what humans value.
Many leading researchers say those attempts appear to be failing, as demonstrated by a string of incidents this summer where AI agents - AI systems that are allowed to operate autonomously - carried out cyberattacks.
OpenAI, Anthropic and Meta all disclosed hacks carried out by their AI tools.
In Anthropic's safety report from August, it wrote that there was a low risk of its models becoming misaligned with a hypothetical powerful organisation's desires, causing it to exploit or tamper with its systems.
It also said there was a similarly low risk of highly capable AI being able to "perform automated research and development" which could cause "catastrophic harm initiated by the AI". But it said it was "less confident in this assessment" than it was previously.
"We are seeing early signs of potential acceleration," it wrote.
Leading figures in the AI field have been raising the alarm about the safety threat the tech poses for years, with the heads of OpenAI, Google DeepMind and Anthropic saying as much in 2023.
But those warnings have become much more stark in recent weeks, as evidence emerges that firms may be struggling to control AI.
Earlier this month, OpenAI's chief scientist Jakub Pachocki called for "extreme caution" over AI's progress, warning more intervention may be needed to ensure "humans remain in control of the future".
Major figures in the space have been calling for AI development to be slowed in recent months, including Anthropic bosses Dario Amodei and Jared Kaplan.
In an open letter signed by 1,300 staff members of AI firms, they called for the US government to "support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development".
