As AI entered a groundbreaking era in 2023, the journey is far from over. Brace yourself for 2024, where a wave of unsettling innovations, possibly featuring artificial general intelligence and hyper-realistic deepfakes, is on the horizon. Get ready for the next chapter in AI evolution!

Artificial Intelligence (AI) has a long history spanning decades, but this year marked a significant breakthrough, especially with OpenAI’s ChatGPT bringing practical AI to the forefront for the general public. Despite its recent success, AI has had a checkered past, with today’s advancements following a trail of earlier failed experiments.

In the realm of AI, advancements predominantly show promise in enhancing areas such as medical diagnostics and scientific exploration. For instance, a specific AI model can assess the likelihood of developing lung cancer through the analysis of X-ray scans. Amidst the COVID-19 pandemic, scientists developed an algorithm capable of diagnosing the virus by discerning subtle nuances in individuals’ cough sounds. Moreover, AI has played a role in crafting quantum physics experiments that surpass human conceptualization.

Not all AI innovations are benevolent. From lethal drones to AI posing threats to humanity’s future, here are some of the most unsettling AI breakthroughs anticipated for 2024.

Age of Artificial General Intelligence

The circumstances surrounding the dismissal and subsequent reinstatement of OpenAI CEO Sam Altman in late 2023 remain unclear. However, amidst internal upheaval at OpenAI, speculations emerged about an advanced technology named Q* (pronounced Q-star) that could pose a threat to humanity’s future.

As per Reuters, this OpenAI system might represent a groundbreaking manifestation of artificial general intelligence (AGI). While details about this enigmatic system are scarce, if the reports hold true, it could elevate AI capabilities to a whole new level.

AGI, or Artificial General Intelligence, represents a theoretical tipping point often referred to as the “Singularity,” marking the moment when AI surpasses human intelligence. While current AI iterations fall short in areas where human capabilities shine, like context-based reasoning and authentic creativity, most AI-generated content essentially reproduces the data it was trained on in some manner.

While AGI has the potential to excel in specific tasks more efficiently than many individuals, scientists caution about its dual nature. On one hand, it could be a valuable asset, enhancing job performance. On the other hand, there’s a concerning aspect as it could be weaponized, leading to the creation of more potent pathogens, orchestrating large-scale cyber attacks, or facilitating mass manipulation.

The concept of AGI has predominantly resided within the realms of science fiction, and numerous scientists remain skeptical about reaching such a juncture. The notion of OpenAI possibly attaining this tipping point may be surprising but not entirely implausible.

As early as February 2023, Sam Altman was laying the foundation for AGI, as outlined in a blog post. Moreover, industry experts, including Nvidia’s CEO Jensen Huang, foresee a potential breakthrough, with Huang stating in November that AGI could be achievable within the next five years, as reported by Barrons. The question lingers: Could 2024 mark the watershed moment for AGI? Only time will reveal the answer.

Hyperrealistic Deepfakes in Election Rigging

A looming cyber threat revolves around deepfakes, which are entirely synthetic images or videos of individuals that can potentially misrepresent, incriminate, or bully them. While AI deepfake technology hasn’t posed a significant threat so far, there are indications that this might be on the verge of changing.

AI has reached a point where it can produce real-time deepfakes, meaning live video feeds that have the potential to be so convincingly realistic that distinguishing between what’s genuine and what’s fabricated becomes increasingly challenging. A study published in the journal Psychological Science on Nov. 13 revealed the emergence of “hyperrealism,” where AI-generated content is more likely to be perceived as “real” compared to actual, authentic content.

The advancements in AI, particularly in generating hyper-realistic deepfakes, pose a significant challenge as it becomes increasingly difficult for individuals to discern between reality and fabrication with the naked eye. While tools to detect deepfakes do exist, they are not yet widely adopted in the mainstream. For instance, Intel has developed a real-time deepfake detector called FakeCatcher, which utilizes AI to analyze blood flow. However, its effectiveness has yielded mixed results, as reported by the BBC.

As generative AI becomes more sophisticated, a concerning scenario emerges where individuals may deploy deepfakes with the intention of influencing elections. The Financial Times (FT) reported that Bangladesh is preparing for a January election potentially disrupted by deepfakes.

With the U.S. gearing up for the presidential election in November 2024, there’s a looming possibility that AI and deepfakes could impact the outcome of this crucial vote. UC Berkeley is actively monitoring AI usage in campaigns, and NBC News highlighted that many states lack the necessary laws or tools to address a potential surge in AI-generated disinformation.

AI-Powered Killer Robots in the Mainstream

Governments globally are integrating AI into military tools, raising concerns about responsible use. On Nov. 22, the U.S. government revealed that 47 states had backed a declaration on the responsible use of AI in the military, initially introduced at The Hague in February.

The necessity for such a declaration arises from the genuine and alarming possibility of “irresponsible” use. Instances include reports of AI drones purportedly autonomously targeting soldiers in Libya without human intervention.

In military applications, AI plays a crucial role in recognizing patterns, self-learning, making predictions, and generating recommendations. The ongoing AI arms race suggests that in 2024, we can expect AI not only in weapons systems but also in logistics, decision support systems, and research and development.

As seen in 2022, AI demonstrated its capability by generating 40,000 novel chemical weapons. Various branches of the U.S. military have invested in drones equipped with AI for superior target recognition and battle tracking compared to human capabilities. Israel also utilized AI during the latest Israel-Hamas conflict, enabling rapid target identification at least 50 times faster than human speed, as reported by NPR.

A highly concerning area of development is the emergence of lethal autonomous weapon systems (LAWS), commonly referred to as killer robots. Prominent figures like Stephen Hawking in 2015 and Elon Musk in 2017 have sounded alarms about the dangers posed by killer robots. Despite these warnings, the technology has not yet materialized on a widespread scale.

Despite previous concerns, there are unsettling indications that this year may witness significant advancements in the development of killer robots. For example, reports suggest that in Ukraine, Russia deployed the Zala KYB-UAV drone, capable of recognizing and attacking targets autonomously. Australia has also introduced Ghost Shark, an autonomous submarine system slated for mass production, as reported by the Australian Financial Review.

The global surge in AI spending is another notable factor, with China escalating its expenditure from $11.6 million in 2010 to $141 million by 2019, signaling a competitive AI arms race with the U.S., according to Datenna and Reuters. These combined developments suggest the onset of a new era in AI warfare.