You can have information, without meaning. The Chinese Room argument considered by Searle and others aims to illustrate and grapple with this.
Certainly you can have intelligence without self-awareness. You might look at how chess for humans was considered a source of transferable skills, utilised for military education from India where the game was created, to Vikings (see 12thC Lewis chessman, & the game tafl). Alpha Go can't do that. Algorythms can be repurposed like the mine-searching one used by Rhumba. Alpha Go can be too in principle, though I can't find any examples. It seems more a proof of principle.
The bigger picture, is about the problems of model evaluation where machine learning is involved - https://en.m.wikipedia.org/wiki/Machine_learning#Model_assessments And more widely, about whether supervised learning can ever form distinct intelligence, or only reflect the supervisors'. Human minds seem to seek simplifying insights from complex data, and to pursue simplification and unification of premises. Machine learning can only pursue a kind of 'metis', a heuristic process of data mining. So far.
There is something of a schism, between those who hold minds to have some capacity to move at least by degree between unfreedom and more-free (we all have physics and biology, culture language etc limiting freedom in some ways, though also enabling it.. ), and those including most AI researchers who tend to see minds as essentially deterministic, and fundamentally interchangeable and intergoperable (universal Turing Machines and the Turing Comleteness of languages), and so seeing free will as an illusion. Steven Pinker sees the 'magic' difference as coming from quantum processes in the brain. Dennett sees free-will and qualia as resulting from evolutionary usefulness (compatibalism). Bostrom see the whole universe as a mind-process with inherent subjectivity (this is embedded in the simulation argument).
There is a kind of middle way, admittedly rather tentative in the rigour with which it's been established, in Hofstadter's Strange Loops model (as discussed on here
Has Hofstadter's concept of strange loop been given a more formal treatment (by him or anyone else) than "GEB" and "I am a strange loop"?). Strange Loops are hierarchical recursive structures with any looping around in the hierarchy, back on itself, creating a tangled hierarchy. As a model for how consciousness works this can address a number of difficulties. Consciousness can be viewed as building gradually with complexity of self-reference, as we in fact find. It can be understood as both an internal subjective experience -symbolising oneself in internal discourse with implications for identity and decisions- but also having external aspects as we in fact find, illustrated by the Private Language argument. Strange Loops seem to offer a way to describe degree of free will, through examining the degrees of feedback which inhibit 'unconscious', material or non-Strange Looping causes.
Buddhist ideas are about understanding 'unshakeable liberation' from causes and conditions, the maximizing version of 'awakeness', and reordering behaviour through a recursive practice (the eightfold way) which has a, I think paradigm-changing rather than transcendental aim. This is pictured as 'crossing the stream' and then leaving the boat, or ascending a ladder and then letting go of the ladder, and in Japan as using vines to unravel vined(intertwining with the teaching symbolised by long-lived wisteria, found at all temples, and bad habits symbolised by kudzu vines, an invasive weed). Also in Nagurjuna's discourse in which even emptiness is found to be empty of inherent nature https://www.google.co.uk/amp/s/absoluteirony.wordpress.com/2014/09/17/nagarjuna-nietzsche-rorty-and-their-strange-looping-trick/amp/ In this way Buddhism can be seen as a practice to maximise 'freeness', not of an ego self but as a how to occupy the stage our self-history has set, not what but how to be moment by moment, through self-reflective 'recursive' practices like meditation and directed contemplations, eightfold way, and four noble truths (interpreted like Stephen Batchelor as directives for the mind rather than ontology). This suggests free 'will' is possible not only by degree but radically, as a different paradigm.
