Anthropic releases 'safe' Mythos AI model at double the cost
Critics question whether prior safety claims were a marketing strategy as selective access becomes industry standard
Anthropic has released a "safe" version of its Mythos artificial intelligence system, introducing enhanced guardrails and user limitations. The new iteration is priced at twice the cost of the company's previous flagship system. This release follows reports that an unauthorised group had accessed the original Mythos model.
The company previously claimed that the original Mythos model was too dangerous to release to the public. However, critics suspect that this refusal was a marketing strategy, noting that selective access has become a key approach for artificial intelligence laboratories. The extent to which the new "safe" version mitigates the risks identified in the unauthorised access incident remains unclear.
Anthropic’s move comes as the broader artificial intelligence sector grapples with questions of safety and accessibility. While the company positions the new system as a responsible release, the significant price increase and restricted access have fuelled debate about the motives behind such safety protocols.
The incident highlights the growing challenges AI developers face in securing their models. As laboratories compete to release advanced systems, the balance between innovation and security remains a contentious issue, with some observers viewing safety claims as a means of controlling market entry rather than a genuine risk assessment.
Despite the concerns surrounding AI safety, recent analysis of US labour data suggests that the technology has not yet had a large-scale impact on the job market. Unemployment in occupations most exposed to AI is actually lower than in less-exposed jobs, with no signs of a significant shift into manual labour roles.
The release of the safe Mythos model adds to a busy period for the artificial intelligence industry. Other recent developments include the unveiling of a new AI model by DeepSeek and NASA’s announcement of plans for three uncrewed missions to the Moon this year.
As Anthropic moves forward with this new version, the industry will be watching to see if the enhanced guardrails prove effective. The company’s strategy of selective access continues to draw scrutiny, with critics arguing that such measures may serve commercial interests as much as safety concerns.
