In recent years, the emergence of "killer robots," also known as lethal autonomous weapons, has sparked significant concern. While these technologies are far from achieving superintelligence—resembling more the character WALL-E than The Terminator—the integration of artificial intelligence into military applications has escalated dramatically, primarily fueled by substantial government funding from the United States, China, and Russia.

Prominent figures such as Elon Musk and Stephen Hawking have voiced their worries about this trend. In 2018, they, along with others, endorsed the "Lethal Autonomous Weapons Pledge," advocating for a worldwide prohibition on such weapons. This pledge describes these technologies as "dangerously destabilizing" and "powerful instruments of oppression." Additionally, the Future of Life Institute organized an even broader initiative in 2017, explicitly warning against the potential onset of a global arms race in lethal autonomous weaponry.

This article aims to highlight the ineffectiveness of such activism, however noble it may be. The reality is that the AI arms race is already in progress, having been underway for years. As we reach 2019, it is crucial to recognize that Pandora's Box has been opened. We must evaluate the existing evidence of weaponized AI and reflect on the geopolitical motivations that have led to this complex situation.

Ready, Set, Gone

To understand the countries leading the AI arms race, we might start with Russia. The Russian government, known for its resistance to established global standards, is firmly opposed to creating any new norms regarding autonomous weapons. Russia has even successfully lobbied to minimize the frequency of meetings dedicated to discussing this issue. They argue that restrictions on the development of autonomous weapons are unwarranted, claiming that few such systems have been created.

However, Russia is actively working to contradict this assertion. Leaked budget documents for 2019–2021 indicate that state funding for AI research and development is nearly set to double from the previous $490 million USD. Russian military leaders, including General Gerasimov, have made it clear that automation will play a vital role in future conflicts, stating that "robots will be one of the main features of future wars."

For insight into potential future warfare scenarios, consider the announcements from Kalashnikov, a leading arms manufacturer in Russia, which is developing a series of autonomous weapons equipped with neural networks that allow them to identify and engage targets independently. Another manufacturer, Degtyarev, has created an autonomous drone named Nerekhta, designed to stealthily approach targets before detonating to destroy fortifications or enemy tanks.

It's important to note that these examples represent only a small fraction of Russia's ambitious autonomous weapons program, which also includes autonomous nuclear submarines and advanced missile systems. It is evident that Russia views its autonomous weapons initiatives as essential to national security, which explains Putin's aggressive stance against the idea of a global AI arms prohibition.

Surprisingly, the positions of Russia, the United States, and China on autonomous weapons are quite similar. Despite their often contentious relationship, these two countries have displayed remarkable unity in opposing any restrictions on such weapons.

What drives this alignment? The answer lies in a staggering figure: the Pentagon has earmarked $9 billion for military AI, specifically to stay competitive with Russian and Chinese advancements. Although this budget is only a fraction of overall defense spending, it has significantly increased in recent years.

Unlike other nations, the United States has already begun deploying autonomous vehicles in combat zones, demonstrating their capabilities in real-world situations. Autonomous naval vessels are now patrolling the South China Sea, with larger systems on the horizon. Additionally, American drones have been used extensively in operations in Afghanistan and Pakistan. It is clear that the U.S. is not interested in halting the arms race; rather, it aims to dominate it.

China's stance appears more complex. In April of last year, it expressed support for a ban on the battlefield use of autonomous weapons. However, the very same day, the Chinese air force unveiled plans for a new autonomous drone featuring "intelligent-swarm" technology. Critics argue that China may be gaining a competitive edge in the arms race, despite its public support for a ban.

Elsa Kania from Lawfare suggested that China is employing a policy of "strategic ambiguity" regarding military AI, publicly aligning with human rights concerns while continuing to develop sophisticated lethal autonomous systems. After all, China's official plan prioritizes becoming a leader in AI, including military applications, often disregarding human rights issues when strategic interests are at stake.

In summary, the three largest military powers have:

• Recognized the importance of autonomous weapons in their military strategies.
• Increased their investments in the development of these technologies.
• Framed the necessity for advanced weaponry in light of advancements made by other nations.

This clearly constitutes a serious arms race.

A Reckoning for Activists

In my view, ongoing efforts to prevent the development of lethal autonomous weapons are no longer feasible, given the current landscape. The well-meaning endeavors of Elon Musk and the Future of Life to forestall an arms race in military AI have been unsuccessful. Currently, Russia, China, and the U.S.—along with numerous smaller nations—are significantly invested in lethal autonomous technologies, viewing them as essential for future military engagements. It is likely that investments in military AI will continue to rise, with states deploying more advanced drones.

This behavior from major powers aligns with their rational incentives, rather than any fictional narrative. For any of these nations, unilaterally supporting a ban on AI weapons would be illogical, as it opens the door for others to gain an insurmountable advantage in weaponized AI. In future conflicts, the effectiveness of autonomous drones will largely depend on the sophistication of their software and autonomy. Machines capable of independent thought and action will have the upper hand against slower systems and human operators.

Given this geopolitical context, what strategies remain for those concerned about the rise of autonomous weapons? One approach may be to engage with the scientists and engineers behind military AI development instead of focusing solely on the states involved. Last year, thousands of engineers protested Project Maven, a Pentagon initiative for military AI, leading to its cancellation. Many of AI's brightest minds may not wish to apply their talents to warfare.

Another potential avenue for activists is to pursue a feasible international agreement that garners support from major powers. While a comprehensive ban may be unrealistic, a more gradual agreement, such as an AI Code of Conduct governing the deployment of autonomous weapons, could be attainable. Countries have a vested interest in establishing shared rules or mechanisms to address diplomatic issues arising from the use of new technologies.

The situation may not be as dire as it seems. Despite the ambitious AI projects discussed here, one expert estimated that the intelligence of current autonomous military machines is akin to that of a young toddler. Although Musk and Hawking may not have succeeded in preventing the AI arms race, their worst fears have yet to materialize.