Kampuni ya ujasusi ya akili mnemba, Anthropic yenye makao yake Marekani inatafuta kuajiri mtaalamu wa silaha za kemikali na vilipuzi vyenye nguvu nyingi, katika jaribio la kuzuia "matumizi mabaya" ya programu zake.
Kwa maneno mengine, inaogopa kwamba zana zake za ujasusi bandia zinaweza kumwambia mtu jinsi ya kutengeneza silaha za kemikali au zenye mionzi, na inataka mtaalamu wa kuhakikisha kwamba uzio wake wa usalama una nguvu ya kutosha.
Katika chapisho la LinkedIn linalotafuta wafanyakazi, kampuni hiyo ilisema waombaji lazima wawe na uzoefu wa angalau miaka mitano katika "silaha za kemikali na au ulinzi wa vilipuzi," pamoja na uzoefu wa "vifaa vya kutawanya mionzi," pia vinajulikana kama mabomu machafu.
Kampuni hiyo iliiambia BBC kwamba jukumu hilo ni sawa na kazi ambazo tayari imeunda katika sekta nyingine nyeti. Anthropic sio kampuni pekee ya akili bandia inayofanya mkakati huu.
Kazi kama hiyo pia inatangazwa na OpenAI, mtengezaji programu ya akili mnemba ya ChatGPT. Kwenye ukurasa wake wa kazi, inaorodhesha nafasi ya mtafiti kuhusu "hatari za kibiolojia na kemikali," akiwa na mshahara wa hadi $455,000, ambayo ni karibu mara mbili ya kile Anthropic inatoa.
Lakini baadhi ya wataalamu wana wasiwasi kuhusu hatari za mbinu hii, wakionya kwamba hii inaweza kutoa taarifa kwa zana za AI kuhusu silaha hata kama wataagizwa kutozitumia.
"Je, ni salama kutumia mifumo ya AI kudhibiti taarifa nyeti kuhusu kemikali na vilipuzi, ikiwa ni pamoja na mabomu ya makundi na silaha nyingine zenye mionzi?" alisema Dkt. Stephanie Hare, mtafiti wa teknolojia na mtangazaji wa kipindi cha TV cha BBC "AI Decoded".
"Hakuna makubaliano ya kimataifa au sheria nyingine inayodhibiti aina hii ya kazi na matumizi ya AI katika silaha hizi. Kila kitu kinatokea kwa siri."
Makampuni ya AI yamekuwa yakionya mara kwa mara kuhusu hatari kubwa zinazoweza kutokea kutokana na teknolojia yao, lakini hakukuwa na jaribio la kupunguza kasi ya maendeleo yake.
Suala hili limesababisha kuwepo kwa tahadhari kutoka kwa serikali ya Marekani, ambayo inatoa maonyo kwa makampuni ya AI huku kukiwa na vita vyake na Iran na shughuli za kijeshi nchini Venezuela.
Wasiwasi wa wataalamu
Pata habari za kina kutoka BBC News Swahili, moja kwa moja kupitia WhatsApp.
Bonyeza hapa kujiunga
Mwisho wa Whatsapp
Hata hivyo, baadhi ya wataalamu wana wasiwasi na hatari za mbinu hii, wakionya kwamba inaweza kutoa taarifa kwa zana za AI kuhusu silaha hata kama wataagizwa wasitumie.
"Je, ni salama kutumia mifumo ya AI kushughulikia taarifa nyeti kuhusu kemikali na vilipuzi, ikiwa ni pamoja na mabomu ya makundi na silaha nyingine zenye mionzi?" alisema Dkt. Stephanie Hare, mtafiti wa teknolojia na mchangiaji wa programu ya AI Decoded ya BBC.
"Hakuna makubaliano ya kimataifa au sheria nyingine inayodhibiti aina hii ya kazi na matumizi ya AI katika aina hii ya silaha. Yote haya yanatokea nyuma ya milango iliyofungwa."
Kampuni za AI zimeonya mara kwa mara kuhusu hatari zinazoweza kutokea kutokana na teknolojia yao, lakini hakujakuwa na jaribio la kusimamisha maendeleo yake.
Anthropic inachukua hatua za kisheria dhidi ya Wizara ya Ulinzi ya Marekani, ambayo iliitaja kampuni hiyo kama "tishio" baada ya kampuni hiyo kusisitiza kwamba mifumo yake haitatumika kwa silaha zinazojiendesha au ufuatiliaji wa Wamarekani.
Mwanzilishi wa Anthropic Dario Amodei aliandika mnamo Februari kwamba hakufikiri teknolojia hiyo bado ilikuwa nzuri vya kutosha na ilihitaji kutumika kwa madhumuni haya. Ikulu ya White House imesema jeshi la Marekani halitadhibitiwa na makampuni ya teknolojia.
Ishara hii hatari inaiweka kampuni ya Marekani katika mashua moja na kampuni ya mawasiliano ya simu ya China Huawei, ambayo hapo awali ilikuwa imeorodheshwa kwenye orodha mbaya kutokana na wasiwasi wa usalama wa taifa.
OpenAI ilisema inakubaliana na msimamo wa Anthropic, lakini bado ilikuwa ikijadili mkataba tofauti na serikali ya Marekani, ambao ilisema bado haujaanza.
Msaidizi wa AI wa Anthropic, anayejulikana kama Claude, bado haijaondolewa sokoni na kwa sasa imeunganishwa katika mifumo inayotolewa na Palantir, ambayo Marekani inatumia katika vita vya Marekani na Israel na Iran.
Facts Only
Anthropic, a U.S.-based AI company, is hiring a weapons expert with experience in chemical weapons and nuclear proliferation.
The job requires at least five years of experience in "chemical weapons and/or nuclear proliferation" and knowledge of "radiological dispersal devices" (dirty bombs).
OpenAI is also recruiting a researcher for "biological and chemical risks," offering a salary up to $455,000.
Dr. Stephanie Hare, a technology researcher, questions the safety of using AI to manage sensitive weapons data.
The U.S. government has warned about AI risks, with Anthropic suing the Defense Department after being labeled a "threat."
Anthropic's AI assistant, Claude, is integrated into systems used by the U.S. military and allies like Israel.
OpenAI is in separate negotiations with the U.S. government regarding AI regulations.
No international agreements exist to govern AI's role in weapons development.
Executive Summary
Full Take
The strongest version of this narrative is that AI companies are proactively addressing existential risks by hiring experts to prevent misuse of their technology. This is a credible response to legitimate concerns about AI's potential to enable weapons proliferation. However, the lack of international oversight and the involvement of military contractors like Palantir raise questions about accountability. The pattern here resembles **ARC-0043 Motte-and-Bailey**: companies frame their actions as "safety measures" (motte) while operating in a regulatory vacuum that allows military applications (bailey). The root cause is the tension between rapid AI advancement and governance lag, echoing historical arms races where innovation outpaces ethical constraints.
The implications are profound. If AI systems are trained on weapons data, even for defensive purposes, they could become targets for adversarial exploitation. Who benefits? AI firms gain credibility while maintaining military contracts. Who bears costs? Global security, as unregulated AI could lower the barrier to weapons development. Second-order consequences include normalized AI-military integration and potential erosion of public trust in AI ethics.
Bridge questions: What safeguards would make AI weapons research truly transparent? How might adversarial nations exploit these systems if they fall into the wrong hands? Would stricter regulations stifle innovation or prevent catastrophe?
Counterstrike scan: A coordinated influence campaign would amplify fears of AI weapons to justify expanded military funding or surveillance. The actual content does not match this pattern—it presents legitimate concerns without overt manipulation. However, the absence of dissenting voices (e.g., AI ethicists advocating for bans) suggests a framing gap worth noting.
Sentinel — Human
The article shows strong signs of human authorship, with direct sourcing and minor stylistic quirks, though some repetitive phrasing suggests possible light editing or translation.