Scheduling meetings is an ability both Siri and Google Assistant possess; however, their current lack of social understanding does not allow them to prioritize these appointments independently.
AI is powerful, but a shortage of social capabilities hampers its potential. Researchers based in China have outlined this belief.
The groundbreaking advancements made in recent times with Artificial Intelligence have greatly influenced our society and daily lives, according to Lifeng Fan of the Beijing Institute for General Artificial Intelligence (BIGAI).
The next key challenge for Artificial Intelligence (AI) in the future is Artificial Social Intelligence (ASI). We argue that ASI will be the next major milestone in AI.
The CAAI Artificial Intelligence Research paper outlined multiple subfields of ASI, including social perception, theory of Mind (wherein individuals recognize others think from a different point-of-view), and social interaction.
Fan asserted that cognitive science and computational modeling could uncover the differences between AI and social intelligence in humans. This insight into current issues and future directions will assist in pushing artificial intelligence forwards.
Fan says:
“ASI is distinct and challenging compared to our physical understanding of the work; it is highly context-dependent,”
“Here, context could be as large as culture and common sense or as little as two friends’ shared experience. This unique challenge prohibits standard algorithms from tackling ASI problems in real-world environments, which are frequently complex, ambiguous, dynamic, stochastic, partially observable and multi-agent.”
ASI must be able to interpret the subtle nonverbal cues such as rolling of eyes and yawning toto comprehend other agents’ mental states, e.g., beliefs and intentions, for collaboration on a shared task to be successfully achieved.
Adopting a more global approach, similar to how humans interact with the world around them, is proposed by Fan as the best way to design Artificial Super Intelligence (ASI). This entails providing an open and interactive atmosphere and exploring how to inject AI models with human-like inclinations.
Fan went on to say:
“To accelerate the future progress of ASI, we recommend taking a more holistic approach just as humans do, to utilise different learning methods such as lifelong learning, multi-task learning, one-/few-shot learning, meta-learning, etc.,”
“We need to define new problems, create new environments and datasets, set up new evaluation protocols, and build new computational models. The ultimate goal is to equip AI with high-level ASI and lift human well-being with the help of Artificial Social Intelligence.”
While many challenges still exist to overcome, the potential benefits of developing AI with social skills are too great to ignore. By teaching machines how to understand and interact with humans on a deeper level, we can create a more harmonious and productive relationship between humans and machines, ultimately improving how we live and work.
Source: Gadgets Now
