“AI is one of the most important things humanity is working on.” – Sundar Pichai

This phrase is a well known phrase from the CEO of Google Inc. He later explained that just like with any other innovation, AI is very helpful but we have to come to terms with it. For mankind it has always been hard to relate to something unknown or new. With AI and the usage it is the same. The broad public hardly knows anything about AI besides the fact that it is existing. The uncertainty might come from the issue of not having a proper definition for it. Many human tasks relate to a certain amount of ‘intelligence’, such as writing and reading an article, engaging in logical thinking or being able to understand speech. Nowadays we use ‘artificial intelligence’ to reproduce these common sense tasks and even more complicated ones with programs. There are for example programs which are able to solve mathematical equations or understand human speech. (cf. Nilsson N.)

AI in our everyday life

We often use AI without realizing it – for instance in every picture search or whenever we use a translator application. Having these advanced applications which almost flawlessly translate words or even full articles is due to AI. AI simply makes things ‘smarter’. It is not only an advance in terms of having a smart TV or fridge – AI has the potential to improve our lives, in innovative health care technologies.

the ‘ki’ in ki:elements

Speech is a source of multiple biomarkers. We use AI in our technology to extract biomarkers from speech to detect early stages of neurological and psychiatric diseases. The developed algorithm understands underlying pathologies of speech and is able to draw conclusions from it. The process of ‘speech’ begins with the thought about communication, so the intention to speak. The phonetic information is then processed into phonemes* with a certain prosodic** information in order to form a meaningful sentence. These phonemes and melodies vary with different intentions, depending on the phrase we want to form, if we want to ask a question or make a statement. All the information is passed on to the ‘speech production system’ as a coded command. So energy is formed to pass the lungs and finally reach the vocal tract where the information is decoded (cf. Trujillo F.). Using AI it is possible to analyse the produced sentence and determine through the underlying structures where in the speech building process something might have gone wrong. Since the speech process is initiated in the brain, it is even possible to detect cognitive difficulties the speaker might have. There are multiple application possibilities for this software. It can be used in the early detection of Alzheimer’s or used for the detection of depression via the voice. The speech-based biomarkers can be used for diseases like bipolar disorders, Schizophrenia, Parkinson, MS and Dementia.

Read more about how ki:elements used AI for example in a study for early detection of Alzheimer via telephone here or how we used our software to detect depression from speech samples here.

* a unit of sound that distinguishes one word from another
** rhythm and intonation of language