Tencent AI Chief Warns of Bioweapons Risk
Tencent’s chief AI scientist warns that artificial intelligence will soon be capable of aiding the development of bioweapons, according to a report by The Decoder citing the South China Morning Post. The executive argues that neither fully open nor fully closed models can solve this security challenge on their own, pointing to the difficulty of separating offensive and defensive capabilities in biology.
The debate over open-source versus proprietary AI systems has intensified across the industry. While US firms like Anthropic advocate for closed models to prevent the proliferation of dangerous capabilities, proponents of open models argue they provide essential tools for defense and collaborative research. Industry discussions highlight incidents where developers relied on open-source alternatives when proprietary safety filters proved overly restrictive.
Security researchers note that safety mechanisms on open models remain vulnerable to circumvention techniques such as abliteration, which can strip away trained refusal behaviors. Despite these risks, regulatory frameworks in both the United States and China have yet to establish binding thresholds for the publication of open-weight models, leaving developers to navigate complex trade-offs between security and accessibility.
Based on reporting by the-decoder.com.
