Anthropic has published a new report explaining that it has blocked five cases where its artificial intelligence (AI) models were attempted to be used for laboratory developments that could have led to the manufacture of biological weapons. “These examples illustrate the tasks for which our models are used, and the judgments we must make when assessing whether a particular biological use is dangerous,” the company writes.
Read more Illa calls to preserve coexistence on a Diada marked by the rise of the far right
The report includes other dangerous uses of AI, such as cybersecurity, conventional weapons, and influence operations. It is the fourth text of this kind published by the company. In this report, all the examples occurred this year and were carried out using earlier models, less powerful than the newest in the Fable or Mythos family, which have more sophisticated safeguards: in principle, they do not allow malicious use easily. The actors behind these operations range from groups with government support to mafia gangs or individuals with political motivations.
One of the uses of Claude for conventional weapons is in northern Yemen, targeting some terrorist group that, according to the report, intended to use AI to achieve three types of weapons: “A guided rocket that used a mobile phone chip as a flight computer” and two types of missiles, one with a range of more than 2,000 kilometers. The use of these tools by terrorist groups is one of the most cited as an imminent danger.
Biological weapons are, however, the greatest danger included in Anthropic’s report. In these cases, the company does not name the authors of the queries to Claude, partly because it does not know if their use is truly malicious. These situations occurred in countries where Anthropic does not provide its services, so the actors accessed the Claude program through dark methods, such as resellers or fake accounts.
In one example, Anthropic decided to block the queries because they were made from a military research institute. These queries were from a scientist asking Claude for help to write a grant application to research the chikungunya virus, a mosquito-borne disease. This type of research can be legitimate in vaccine development, for example, but it can also create super-developed versions of pathogens. In this case, the scientist wanted to design mutations in the virus that made it more harmful as it repeatedly infected live animals. “This type of research could be used to develop better vaccines and treatments against the virus, but it could also be used to make the pathogen more dangerous,” the report concludes.
The rest of the cases include research to make avian flu more contagious, to analyze the smallpox virus more resistant to the immune system, or the creation of a catalog of toxins from venomous animals to develop new drugs or possible weapons.
The publication of the report comes a day after a young Anthropic employee resigned and publicly warned on the social network X about the risk of civilizational extinction for humanity that AI could cause in the coming years. One of the options is precisely this malicious use by humans.
This type of report presented by Anthropic is similar to those Facebook and other social networks published a few years ago about the circulation of covert operations or dangerous information on their platforms by malicious actors aiming to undermine institutions and democratic processes. With AI, the potential damage would be much greater.
Read more Christopher Columbus, founder of the yogurt group La Fageda, dies
